my-opencode-projects

Compresses pipeline stage output into minimal, structured inter-agent payloads. Reduces token waste by 50-70% between stages while preserving all actionable information. Use when: between stages, reduce tokens, compress context, stage handoff, agent payload, token budget, optimize pipeline.

Natural language query interface for Azure DevOps. Ask questions about bugs, regressions, velocity, releases, team performance, and work items without writing WIQL. Use when: ADO query, regressions, root causes, velocity, team metrics, release bugs, sprint analysis, bug trends, what broke, how many bugs, find tickets, ADO search, work item analysis, regression report, team Madras velocity, release quality.

Validates that implementation matches the discovery contract. Checks Must Do/Must NOT Do items, guard rails, and cross-repo API consistency. Use when: validate contract, check compliance, verify implementation matches discovery, API contract check, cross-repo consistency.

Get Jenkins build status, summarize failures, identify root causes, and explain what fixed the build. Use when: checking Jenkins build, why did build fail, what broke the pipeline, build summary, CI failure cause, Jenkins job status.

Records patterns and outcomes from completed pipelines to improve future runs. Tracks what worked, what failed, and builds institutional knowledge. Use when: save lessons, record outcome, pipeline completed, what did we learn, improve future runs, pattern library.

Classifies requirements by type and selects the optimal pipeline mode. Analyzes requirement complexity, target repo, and risk to determine whether to use full-feature, quick-fix, hotfix, or refactor mode. Use when: classify requirement, what pipeline mode, how complex is this, should I skip discovery, requirement classification, complexity analysis.

Generates live pipeline progress visualization and status reports. Shows current stage, timing, confidence scores, and overall progress. Use when: pipeline status, show progress, what stage are we at, pipeline visualization, progress report, timing, how long did it take.

Generates a polished, copy-paste ready PR description from a git diff and implementation context. Can be used standalone or as part of the pipeline. Use when: generate PR description, write PR body, create pull request text, PR summary, changelog entry.

Transforms vague user statements into structured, actionable requirements. Handles ambiguity, asks clarifying questions, classifies requirement type, and produces a pipeline-ready requirement document. Use when: new feature request, vague requirement, unclear scope, user says 'I want...', 'we need...', 'can you add...', 'build me...', requirement analysis, scope definition.

Safely reverts all changes made by the current pipeline run. Restores files to their pre-pipeline state. Use when: undo changes, revert pipeline, rollback, cancel changes, start over, something went wrong.

Generates test boilerplate and structure based on implementation report. Pre-generates imports, mocking setup, test class skeleton, and test method stubs. Speeds up the test agent by ~40%. Use when: generate test boilerplate, scaffold tests, create test skeleton, test setup, mock setup.